US2659072A - Data transmission system for digital calculating machines or the like - Google Patents

Description

New. 10, 1953 J. F. COALES ETAL 2,659,072

DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE Filed Jan. 13, 1949 8 Sheets-Sheet 1 ////////1l' T11 m %m, 5 E a :3 m a m .LBZEBBEEE 2a 222 L, m; F|G.2.

T 1? ("E @E' A #[aaafifia arm zzmaa v "2225 5 E522 T 222 5MB I Inc :155 B5 [155 Ba 1, BE EB QB 55 ZIE B E E B E E E E B 5 E La m5 ;'i$ 7 ;It;l Zl3l4-I fi I F v x Inventors: FIG-I- l1 ii Ah #231251 52355? ELLIOTT Normgn D 8 0: Samuel Edwin HER John TINDALE I] Lwaaa'aaaa: I Er E g"..- 5 EB: am; am a an 5 a a a a a T a: w ////m FIGJA.

7 v/@ i By e HM Attorney Nov. 10, 1953 J. F. COALES ETAL 2,659,072

' DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE Filed Jan. 13, 1949 8 Sheets-Sheet 2 John I. COALES l l l a a b" zb P I o a ---a b- -b m -H] A ---m, o

b b 9 FIG.5 --3 CD R Cl. GL, GR

F|G.6. 7720 l PI I P 4' P6 do mo;mo P do P. d: gm p, d| P'o B m I Z f-4 H P /W Pu o-4f m Po W/ u 1:: d

50 a F, M 5 11 D :50 5 P C7 C1 FIGJO William Sidney ELLIOTT Inventors: Norman Douglas HILL Samuel Edwin HERSOI J TINDALE I CZMQL y Attorney Nov. 10, 1953 Filed Jan. 13, 1949 J. F. COALES ETAL 2,559,072 DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE 8 Sheets-Sheet 3 I (22m) D JohrjNDALE fig I2 By Attorney Nov. 10, 1953 Filed Jan. 1:5, 1949 J. F. COALES ET AL DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE 8 Sheets-Sheet 4 PCA. (FIG-l7 A TB. TBC. I TBc2. mm. FIG. 23 $10.25)

f CL! C 7 BD. 1 (Ell-g6) C (F lg le) L2 (Fm-3.22)

Fl6-l FIG l3 d f M P PCA CRT E c (FIG-I7) A FIG. I4 TB I (PIC-3.2!) LDG X ROG (FIGS I812 Flsfs lot 22) J- PTG (Be a) D6 (FIG-l6) D Inventors:

John F; COALES William Sidney ELLIOTT Norman Douglas HILL Samuel Edwin HERS M John Truman M C/QWM tt Nov. 10, 1953 Filed Jan. 13, 1949 CRT J. F. COALES ET AL DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE 8 Sheets-Sheet 5 FL ME 6 PC A SB FIG.I5. b

B PCA I I (I l i2!) (PTA?) PTG 0G (FIG-l9) (Flam) D Inventors:

John 5. GOALES WillIgm Sidney ELLIOTT Nopmamnouglas HILL Samuel Edwin HERSOH John TINDALE mama;

NOV. 10, 1953 co Es ETAL I 2,659,072

J. F. DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE Filed Jan. 13, 1949 8 Sheets-Sheet 6 :wvmmms: John F Coalea, William 8 Elliott, Norman D Hill, Samuel E Hersom,

Jo Tindale.

J. F. COALES DATA TRANSMI ET AL Nov. 10, 1953 2,659,072

I SSIQN SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE 8 Sheets-Sheet 7 Filed Jan. 13, 1949 m m??? mmm um a. @6 5 mvwgi m nmum m on P. w J l JWNSJ A llkh lmmilll Q lllkn v m|lll I. H W H I H H $9 I @i w +52 mm an 4+ r Nov. 10, 1953 J. F. COALES ETAL 2,659,072

DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE Filed Jan. 13, 1949 8 Sheets-Sheet 8 P76 86 if; 129 76 126 128 Invntors: Uohn F Coulee, Willim 8 Elliott Noman D Hill I; Ber-songs: JohnTindale Patented Nov. 10, 1953 UNITED STATES PATENT OFFICE DATA TRANSMISSION SYSTEM FOR DIGITAL CALCULATING MACHINES OR THE LIKE Application January 13, 1949, Serial No. 70,764

10 Claims. (01. 340-354) This invention concerns data transmission systems for use, for example, with digital calculating and like logical machines for converting input data or information into a resultant output which may be in the form of an electrical signal or of a mechanical motion. The input data or information from, say, a movable member is, at least in part, converted into digital form, i. e. in the form of discrete phenomena or states of devices, or in the form of sequences of discrete events, and means is provided for interpreting the transmitted data in a convenient form, for example as a movement of a controlled part or as a numerical or mathematical result.

The representation of data or information by discrete states of devices means that a device can only represent given data when in a steady state which is definite and distinct from any other such state which the device may assume, and that the relationship between a steady state and the corresponding data or information is unique, i. e. the said data or information cannot be represented by any other steady state of the device. Furthermore, the transition between one steady state and another represents no intelligible information.

As examples of systems employing digital representation of data may be quoted electronic calculating machine in which the data to be operated upon is represented by electrical pulses, which operate suitable devices, the presence of a pulse causing the device to change from one steady state to another and the absence of a pulse leaving the device unchanged. Another example of such a system is the mechanical transmission of information by discrete linear or angular displacements of mechanical elements to one or another of a plurality of definite positions or attitudes.

The data or information may be represented either by the combination at any instant of the simultaneous states of all of a group of separate devices, or by the sequential states of a single device. Thus, for example, where the data is a number in the binary system, or scale of two, in which only the digits and 1 occur, the number 2 in the normal decimal scale corresponds to the number in the binary scale, and may be represented by the simultaneous states of two devices, the first of which is in the l-state and the second in the O-state. Similarly the decimal scale number 3 is represented by both devices being simultaneously in the l-state, and the decimal number 4 by three devices the first of which is in the l-state and the other two in the 0-state. The devices may, alternatively, have three possible steady states, such as the +1 state, the 0 state and the 1 state," or alternatively the full-on, the half-on, and the oil states. Such devices can be used to represent decimal scale numbers in the ternary scale which possesses only the digits 0, 1 and 2. Systems which employ the simultaneous states of different devices are commonly termed parallel systems, and require as many separate channels as there are digits in the highest number to be indicated, whilst systems which employ the sequential states of one device are commonly termed series systems.

In both systems, data is commonly transmitted from an input or source to an output by way of one or more channels in the form of pulses which may have a variety of physical formsusually electrical or mechanical in nature, although pulses of an optical, magnetic or any other nature could be employed if desired. The absence of a pulse, in the ordinary significance of the term, on any significant channel or at any significant instant, may represent a digitsay the digit 0-in the appropriate place in a number, and may also be regarded as the pulse or like phenomenon in the same way as that which represents the digit 1, since it has a discrete and unique interpretation.

From this it follows that, if a digit-say, 0-15 to be represented by the absence of a pulse in a significant place in the pattern, the corresponding marking may be physically indistinguishable from its general backgroundfor example, if the markings appear on an opaque strip or disc, a marking representing the digit 0 is constituted by a significant zone of the surface of the strip or disc which may not be specifically defined by a boundary. The term marking is to be understood as including such a significant, though physically undefined, zone. It will, however, be appreciated that such a marking may be used to indicate some control function. For example, in a binary system of numbers, a non-reflecting opaque strip or disc may carry transparent and reflecting markings representing the digits 0 and 1 respectively, or vice-versa, whilst a non-reflectin zone may be used to indicate an error or some control signal.

In this specification, the term marking will be used to signify any durable discrete phenomenon which is capable of identification and presentation as an event to which a unique interpretation can be assigned. The process of identification and presentation will be termed herein reading, and the term "durable is intended to indicate that the marking is not destroyed or changed by the reading process. An event may be constituted by an electrical pulse or a mechanical displacement or any other perceptible occurrence such as an optical or audibleeffect. A group or train of pulses will be under-- stood to include the case of a single pulse or the absence of a pulse in a single significant place where such a single significant place hasa-definite or discrete interpretation.

The present invention is primarily concerned with the generation of pulse trains which represent, in digital form, the data-to-be transmitted. This data may have any desired significance, but frequently represents the instantaneous position of a movable member, or the; value at a. function such as a trigonometrical ratio of an angular displacement of a rotatablemember.

It is an object of the present invention to provide. a form of generator which is simple and efiicient and by which a largerange of values of a variable can be represented.

It is broadest aspect the invention provides a digital data transmission system comprising a pattern composed of groups of markings, means for reading each marking as a corresponding digital signal and means for interpreting the said digital signals, each group of markings representing a value of a desired function of the position of the pattern relative to the reading means.

Preferably, a marking is constituted by a change in a characteristic of a surface, and the reading means comprises means for detecting and responding to the said change.

The change maybc-of an optical nature, and

the reading means may comprise a light sensitive element responsive to-the-optical' change. Thus, for example, the pattern may comprise a plurality of transparent markings on a generally opaqueground or vice-versa. The readingmeans may then comprise alight source and a light sensitive element located on opposite sidesof the pattern, and means is provided for causing relative movement betweenthe pattern andthe reading means. It Will be clear that the markings may alternatively be of-' areflecting character on a non-reflecting ground, or vice versa; ora com.- bination of reflecting and transparent markings on a light-absorbing ground may be used. The latter arrangement is of particular advantage where the data is represented by digits onthe ternary scale;

The invention will be more clearly: understood from the following descriptions; given by way. ofexample only, of various embodiments thereof, reference being directed to the accompanying drawings in which:

Figure 1 illustratesa pattern of'discrete markings: which: represent on the binary scale the digits to +15;

Figure 1a shows an alternative arrangement of the left hand part of the pattern of'Flgure. 1 for representing a negative number by a. complement;

Figure .2shows the pattern'ofFigure 1 but'with adjacent markings. in each vertical column coalesced;

Figure Sshows the pattern of Figure 1 but with the adjacent markings in eachhorizontal row coalesced;

Figure 4' shows the samepattern but with the markings coalesced in both the vertical and horizontal. directions;

Figure 5 shows a portion of the pattern of Figure 1 having subsidiary markings for correcting the traverse with respect thereto of a scanning light beam forming part of the reading means;

Figure 6 illustrates certain waveforms associated: with the pattern. shown in Figure 5;

Figure '7 illustrates a modification of the pattern of Figure 4 showing an alternative method tothat of Figure 5 for correcting the traverse of. the. scanning, beam with respect thereto;

Figure. 8. illustrates certain waveforms derived from the pattern of Figure 7;

Figureiishowaapattern consisting of two separate sets of markings, each set representing a different mathematical function of the same variable and arranged in the manner of Figure '7 for correcting the traverse of the scanning beam with respect thereto;

Figure lil illustrates certain waveforms derived from the pattern of Figure 9;

Figure 11 is: a simplified block. diagram. of: a typicalcircuit, for an electronic data transmission.

systenn for indicating. at. a: distance the angular position ofv a disc;

Figure 12- is a fragmentary illustration toa largerscale of the discemployed in the arrangemsnt' of Figure 11;

igure 13; isa block diagram of a. circuit dc.- signed for use with apat-tern such as that. shown. in-Figures. 9 and 10;.

Figure 14 is a block: diagram of a circuit. designedfor usewith a. pattern-as shownin Figures- 5 and 6 for preventing the transmission of. a. wrong number;

Figure- 15 is a block diagram ofv a circuit. arrangement for generating a digital output-representing a function oftwoindependent. variables;

Figure 16. illustrates schematically a; possiblev form of markings for use in adata transmission system operating on. the ternary scale;

Figure 17 illustrates a known circuit suitable. for use as the photo-cell amplifier which is. indicated in block form at PCA inv Figs. 11, 13, 1.4. and 15-;

Fig. 18. illustrates a known. circuit suitable for. use as the digit pulse gate which is indicated in. block form atDG in Figs. 11, 13, 14 and 15, and also as the gate shown in. block form atIG in. Fig; 13. with. an appropriate change. in an. input connection;

Fig. 19' illustrates a known circuit suitable for. use as. the.- single-output clock pulse train generator indicated in block form at PTG in Figs. 11 and 15;

Fig, 20-illustrates aknown circuit suitable. for use as the clock pulse train. generatorv having, multiple outputs which is shown in block. term at P'I'Gr in Fig. 13;-

Fig. 21 illustrates a typical known time base cireuitfor use where indicatedinblock form at TB in- Figs. 11, 1.3, 14. and 15;

Fig. 22 showsa known circuit which is Suitable for use as the beam deflection circuit indicated in block form at EU in Fig. 13;

Fig. 23 illustrates a. known circuit suitable. for use as the starting time control circuit indicated in blockform atTBCI in Fig. 13 and also as the scan speed control circuit indicated in block form at TBC2. in Fig. 13, with an appropriate transposition of the. input connections;

Fig. 24 is a suitable known circuit. for use where indicated by thebiock PTG in Fig. 14.

In the following description, the circuit diagramsin block form will first bedascribed asa whole, typical detail circuits of each: block. be

ing later described with reference to Figs. 17 to 24. Each block is identified with the corresponding detail circuit by the reference thereon to the appropriate figure (or figures, where two detail circuits are combined in one block).

Throughout the following description with reference to Fig. 1-24 the data transmission system concerned is of the electronic type in which the data is in the form of binary numbers or "words" which are transmitted as discrete electrical pulses which are derived initially from a pulse train generator. The pulses required to represent a binary number or word, together with any pulses constituting operation instructions, are selected by means of a pulse distributor and gated by pulses from a photo-cell, the illumination of which is varied in accordance with the markings on a relatively movable pattern. The markings are scanned transversely to the direction of relative motion by a light beam.

The general principles of operation will be more clearly understood by reference to Fig. 11 which shows a simplified circuit for transmitting to a receiver (not shown) digital data in the form of binary numbers or words representing the position of a movable element ME such as the rotatable disc which is illustrated fragmentarily on'an enlarged scale in Fig. 12. The disc ME is here assumed to be generally opaque,

and is provided with a pattern p of transparent .51

markings m1 described more fully below which are scanned radially by a light beam SB derived from the fluorescent screen FS of a cathode ray tube CRT by way of a focusing lens FL. The light beam SB falls on the photocell PC, the output from which is in the form of pulses corresponding to the markings and is fed to an amplifier PCA and thence to a digit pulse gating circuit DG. A clock pulse train generator PTG continuously generates pulses of equal magnitude at regular intervals which are fed to the gate DG, those which coincide with the output pulses from the amplifier appearing in the output from the gating circuit DG. Trigger pulses are also derived from the pulse train generator PTG for triggering the time base circuit TB.

Referring to Fig. 1 of the drawings, a number of transparent markings m1, arranged in vertical columns 0 and horizontal rows 1, are applied to an opaque strip or tape ME which is to be traversed in the direction of its length past a scanning position where the scanning light beam SB (Fig. 11) is caused to fall upon the strip. The beam scans in the direction of the arrow 3, and, Where it encounters a transparent marking m1, light is transmitted to the photocell PC (Fig. 11), to cause a pulse in the output circuit. Each intersection of a column 0 and a row r represents a significant place in a binary number to be transmitted, there being as many columns 0 as there are numbers in a pattern, and as many rows r as there are digits in each number.

Each marking m1 represents the digit 1, whilst the absence of a marking in a significant place (indicated, for example, in dotted lines at me) represents the digit 0. marking m1 at a significant place in which, in another number, a marking m1 may appear, has equal significance with the presence of such a marking, and is regarded in this specification as a marking. Only a few of the total number of markings mo in the pattern are indicated on the drawing, as will be understood.

The markings mo, m1 are shown arranged in a series of columns 0 (only a few of which are ac- Thus, the absence of a tually referenced on the drawing). The columns 0 are disposed symmetrically about the axis X-X except that an additional sign marking ms appears at the foot of the vertical columns 0 to the left of the axis X-X. This'sign marking ms is thus traversed by the scanning light beam SB before any column of markings c is scanned, and is used to indicate that the binary number represented by the photocell output is negative. Each column of markings 0 represents on the binary scale a corresponding number of the series 0-15, as shown by the numbers appearing below the right hand portion of Fig. 1. The position of the axis XX represents zero.

It will be seen that in scanning the pattern of Figure l, the least significant digit is transmitted first in time. This arrangement is for the convenience of the later computing stages as will be appreciated by those skilled in the art. In writing numbers in the binary scale, the usual convention of placing the most significant digit first will be adopted throughout this specification.

In a circuit arrangement of a calculating machine employing a pattern of the kind shown in Fig. 1, it would be customary to arrange that a pulse in the photocell output at any instant when the scanning beam traverses the sign marking ms is not recorded as the first digit of a number but serves to indicate, in a later stage, that the subsequent digit pulses represent a negative number. Where the beam sweeps a column to the right of the axis XX, no pulse occurs in the photocell output for the duration of the sign gating pulse so that the subsequent digit pulses represent a positive number.

In certain calculating machines it may be convenient to employ the complement of a number instead of the modulus of a negative number preceded by the negative sign. In this case, a pattern as shown in Figure 1a is used, the sign marking ms then indicating that the subsequent series of digits represents a complement.

Fig. 2 illustrates an alternative form of pattern corresponding to that shown in Fig. l but in which all the adjacent markings representing the same digit value in each column 0 are coalesced to form a single continuous marking mo, m1i. e. the rows 1' are contiguous. In this case, the photocell output operates a gating circuit to permit the passage of an appropriate number of digit pulses (according to the length of a marking 1121), which are selected from a continuous train of clock pulses produced by a clock pulse train generator. This arrangement enables a narrower strip ME to be employed without involving undue complication in the circuit arrangements.

Fig. 3 illustrates a further alternative form of pattern in which all the adjacent markings representing the same digit value in a row r are coalesced to form the markings mu, m1-i. e. the columns 0 are contiguous.

Fig. 4 illustrates a still further alternative arrangement of the markings mo, ml in which both the columns 0 and the rows 1 are made contiguous.

It will be appreciated that if, in scanning the pattern of Fig. 4 a scanning beam starts to traverse along boundaries between columns of markings, owing to the finite size of the light beam and to a possible relative deviation during the course of the scan, outputs from the photocell may be obtained in digit positions where, in fact, no output is required, and vice-versa;

answers ii-iessps rma example. s {th scan a th :msition 3 deviates liehi c .to th l zm s bil t "of ,ceneratin an .output serene-s als e snen ins t the binar :numb r aim whichis nierem y 2 1192 tha to unit- ;It wil b .seen that i this 5m tr i inc a vhqimdar to the :leit a colum representin w pe numb th e ror annot-b reat r than n t. whi e in z-tra-versing .right .hand boundem e fpfisu heolumhs errors. of larger ma ni u .eenia ise- ,fifiiu-f .cn term ref pat e n by e aiiseyc awhi hth c nce of such erron ous cutputs is substantialla reduced or eliminated. In thi mt ern. i hg 1i eit vmath ns m and'mi a .diifi ffite i :t manner il ustrated .inj'ig. .1 .and ha subsidiary-control ma k 1 b ass eiated therewith a di gona y qpposite .corners- Ea h cont o me ismg :a .b of a eng h. measured in ;the.dir ectionof ;scan-as indi ed-by e heearrcw 3 etc .eneehali eth lspcc h be ween adja ent .rbws (whi s i s widthis equa to eh ehalf spacing between adjacent columns 0.

[If th s an s a ts to trave a one," or .durin ytssve e me deviated t a @3911 near the edgerof a column-c thescannh glight beam will stall iiipon one or ,more of the control markings .a, l2,. the Jenn scannedif the de ia n as to right and :thelatter bein scanned if .fihevdchiation is to -th e 1,eft. fIihe resultant phoputput pulses are uti t gate clock guises derived .from the clock pulse train genrator ,to;a{beam-.deflectiqn circuit which applies 8 rcqnrefi iin Jateral deflection potential :to the ;X-p1ates;cf ;the1cathosie-ray tube .in the sen e f r deflecting the scanning-beam towards the centrel ne of the'echazmp heinescanhed. Thechance .91 error in aesultant digital-output thus wsubetahtiaiiyred ee h ielimina .Eis- .-sh ws he WEE-198 Wa o m a so at d zssit sh patter o i The wavefor .CD

th digit v= .s esl imi s w ve orm ,generated :hy the s e k u e t ain s et r. wh l the w ve er m QCL, re res nt .ieian right d flec on ph s w veie ms spect l which also de ived tram h w i Pulse tr n vgem aerator. ;P represents the photocell output-waverwhieh of a lef defl ct on s te P and three edie t a ate @P es 221 in (th sie iiieant pla n h scan li e in- 411M :hy the E ZQ ,31 ga e h seie eu :a P- -$I!he 'resuitanti teat wa e esm isia sho n at CGL havin a s n le .ieft deflect on :Pui at .1 which op rat e t deflect th beam $8 to the .leit. at is whilst the r sulta t ri gate waveform exhibits no deflection pulses and is mt operative o defle th scanning b The resultant di it utput wa ef m is ishownet .D, and on ists of ze o u do at =1ihe first-significant marisinemo d sit outpu pulses d1 at the second, third and fourth signiflcant placesm n The binary number transanitted is thus ;1-1 10,or 14 on the decimalscale.

Fig. -7 shows an alternative method of correcting the position of the scanning beam on a..pat-

this figure, the pattern of markings .mi-is in-the formillustrated inFig. 4 and the greatest chance ,-of Serious .error i outpu Signal occur if :the scanning light beam traverses al the ri ht-hand ed e f alt rnate co umns caches c1 repr senting Qd num r- A-.r. w 1 :0! initiator markin 2 is therefor ;a ans Eb the pattern, the cen re ine .e each mar in forms.

. be hsc mei eh with th right-ham ers .01 column ,01 immediately alcove .it representing an ;o d,d number. :If the scanning beam is positioned to the right of the centre :line of the column c; or markings to be scanned, the photo cell-output-P .(Figna) contains a pulsePr-which serves-to gate an initiator clock pulse 1 fromthe pulse train generator to -.a correctioncircuit vfor deflecting the scanning beam through one col- .umn pitch. This condition is indicated Jay .the arrows, 13a, in the figure. ;If, however, the (scanning beam is located .as .at .4, no .initia O -marking -2 is encountered prior to the scanning of the pattern markings m0, m1, .and hence no .correctionsignal is applied from the photocell to .alterthe position of thebeam.

Fig. 8 shows the initiator pulse 1 which .occurs at the-instant .whenthe beam is traversing .the .row ,rrof initiator .markings. The Waveform -P represents the Output derived from the photocell whehithe scanszthe path 3, 3q,.whi1st the waveform .1511 represents the output ,trom .the photocell when the beam scans :the pathA. The corresponding digit pulse outputs 13, :D1 represent the binary numbers 1:101 wand 10 10 which com'espond to the decimal scale um e :13 and .10 respectively.

"Fig. *9 illu t a es -.-a gpattern consisting 20f two difierent ;sets of {markings .-P 1, 2B2 itespectively to- ;eet er w t timin m rkin 5a, d two ;rows n, mg of initiator markings 2 2b respec- -:tivel y, and (Fig. 1.0 shows the associated wave- In the latter f gure, the waveform ,P represents the photocell output, the waveform ,Crqthe itime base-speed control circuit .waveform, 1 the initiator markings waveform, -C:e the .beam deflection Voltage .SwitQhrOfl waveform, V t beam de e t on volta e w veform. n D :th di it mitm Wave em wh ch is to be en mitted. beam 1513 a ay trav rses -,a timing spankin 16a hr 51b :at .th beginning 'Or send re- :s e ti eiy Q .a-seahhm ep th rm k g a hm ueinc a ea n 11 1 1 PM in athe inhotficell amtnu wh se to orrectin a -.manner to-,-be desorii2ed with -;reference to Fig. 124..th

instant of start :of a s n ng sweep, e g

weer a number cans, w ils them i 51) produces a atin p lse Pa :i th :PhOtocell out- :w 1! which serve t :ee i t c th slic ,o sc n, re ain avera ed over a :numhe eiscans, so ha he team timin i centinuq sly under qcont c zbyrtheipattern Pl, =Pz;itselforder :to effect the desired control of :the start and speed of :the scan w p :i :i :predenied ".130 arrange that :the :pulses Pu and 2a are differentiated, in the time base contro .cir- .cuit Tao .of :Ei 13, so as :tczn duce a sho apositiveegoing pulse at the ,comxnencement of .each of -z-the isaid :pulses and a short .negatifia- :going pulse at the ends thereof. The :tim {clock tpulses Cu and Ca are ,also differentiated in the .same manner, the positive-goin pulses which result :normally appearing :later :in time than those :derived from the :pulses Pu and J PtZ,

; respectively, and the resulting negative-going pulses normally appearing earlier :in time than those derived from the pulses Pu and "Pm, re- :spectively. Under these conditions, the time- :base start .ci-rcuit exercises no correction on "the time-base circuit. If, :however, the start of a beam scan ;is -retarded, the -positive-going pulse derived from the pulse Pu o r Pm will tend to coincide in time with time positivegoing pulse derived tram th P se .Cn e z esnectively where s orresp ndin he ati o ns u se will tend to separate more in time. On the other hand, if the start of a beam scan is advanced, it is the negative-going pulses derived from the pulses Pu and Cu (or Pa and Ca) which tend to coincide in time while it is the corresponding positive-going pulses which tend to move to a greater distance apart in time. The time-base control circuit utilises the resulting preponderance of positive-going signal, or negative-going signal, as the case may be, to shift the beam scan in the direction of scan so that it is advanced or retarded as necessary. These signals are also utilised in the appropriate manner to vary the slope of the time-base wave and thus to speed up or slow down the scanning sweep as required.

After traversing the first timing marking a, the scanning beam traverses a first row T1 of initiator markings 2a. These markings are arranged in the same manner as the initiator markings 2 in Fig. 7 to be encountered by the scanning beam SB should it lie on a path where there is a chance of error arising in the photocell out- Put P.

Such a path is indicated by the arrow 3 in Fig. 9, and the markings 2a are operative to cause a deflection of the scanning beam. In the example shown, the deflection is approximately one column pitch, and is to the left.

If, as shown, the beam then strikes one of the initiator markings 2b in the second row r2, another initiation gate pulse PiE is produced in the photocell output which gates a second initiator clock pulse C12. This pulse goes to the beam deflection circuit which then applies a second left deflection bias voltage V2 to the X-plates of the cathode ray tube to deflect the scanning beam SB through approximately one column pitch to the left as shown at 30, this voltage V2 being maintained, as before, for the period of scan of the second pattern P2. The voltage V2 is again removed after this pattern has been scanned.

The provision of two rows T1, n of initiator markings in the pattern of Fig. 9 may be thought desirable as a precaution against relative displacement of the scanning beam and the pattern during the period of scan of pattern P1. It will be appreciated that a multiple pattern such as that shown in Fig. 9 may comprise individual patterns representing more than two functions, and each may contain any desired number of digits. It may, however, not always be necessary to interpose initiator markings such as 2b between adjacent patterns.

Figure 13 shows in block diagram form a circuit which embodies the principles of operation I described above in connection with Figures 9 and 10. A cathode ray tube CRT is connected to a time-base circuit TB so that the spot on the fluorescent screen FS of the tube is deflected across the screen at uniform speed in a straight line. The scanning beam SE is focussed by a lens system FL on to a movable pattern of markings ME arranged in any desired way-for example, as shown in Fig. 9. The time-base voltage is synchronised by trigger pulses received from a clock pulse train generator and distributor circuit PTG, whilst the output from the photocell PC is fed to an amplifier PCA whose output P is in turn fed to a digit pulse gating circuit DG. The clock pulse trains generated by the circuit PTG are also distributed to the digit pulse gating circuit DG so that the output D (Fig. therefrom is in the form of discrete digit pulses do, d1 whose sequences correspond to respective 10 sequences of the markings scanned by the beam SB, and represent on the binary scale the numbers l3 and 7 when the scanning beam sweeps the path 3, 3a, 3b, 3c of Fig. 9.

The pulse train generator PTG includes a known form of pulse distributor circuit having a plurality of output channels, one for each possible significant place on the pattern. The distributor executes a complete cycle of distribution operations for each sweep of the scanning beam SB. Some output channels from the distributor are commoned, as for example the channels feeding the digit gate BS and the channels feeding the initiator gate 1G. The initiator gate IG receives the output waveform P (Fig. 10) from the amplifier PCA which may include initiator gate pulses--as shown in Pu and PiZ in Fig. l0occurring at instants when the distributor feeds clock pulses to the initiator gate. When this occurs, the initiator gate IG passes a clock pulse C11 from the distributor to a beam deflection circuit BD which applies a left deflection voltage V1 to the X-plates of the cathode ray tube. This voltage is maintained for the full width of the pattern P1. At the end of the period of scan of this pattern by the beam SB, the beam deflection circuit BD receives a clock pulse CB1 from the distributor and removes the beam deflection voltage V1 so that the beam is returned at 3b to its original path.

The output P from the amplifier PCA is also fed to the time-base start control circuit TBC1 and to the time-base speed control circuit TBCz, the outputs from both being connected to the time-base circuit TB. The start control circuit TBCl serves to alter the longitudinal shift on the scan deflection and is operated if the scanning beam SB falls on the first timing marking 5a at the incorrect instant. The speed control circuit TBC2 determines the slope of the timebase wave, and so serves to speed up or slow down the scanning sweep according to Whether, on the average of a number of scans, the scanning beam SB strikes the final timing marking 5b early or late, as already described.

Figure 14 shows in block diagram form a circuit for use with a pattern of the kind shown in Figures 5 and 6. In the diagram, the cathode ray tube CRT, the movable pattern ME, "and the photocell PC are shown as a single block for convenience of drawing.

The output from the photocell amplifier PCA is fed to a digit gate DG and serves to gate digit pulses received from the pulse train generator PTG in accordance with the digit markings m1 in Figure 5. When the path indicated by the arrows 3, 3a in Figure 5 is being scanned, the binary number is generated and appears in the output from the digit gate as the waveform D (Fig. 6) the space do indicating the digit 0 and the pulses d1 indicating the digit 1.

The output from the photocell amplifier PCA is also fed to a left deflection gate LDG and a right deflection gate RDG. The left deflection gate LDG receives a train of pulses indicated in Figure 6 by the waveform CL whilst the right deflection gate receives a waveform as shown at CR. The waveform Cr. consists of pulses occurring at regular intervals at the beginning of each period between successive pulses in the digit clock waveform shown at CD whilst the waveform Ca consists of similar pulses occurring towards the end of each period between successive pulses in the waveform CD.

If the scanning beam SB strikes a left deflection marking a on the pattern of Figure 5, a gating pulse occurs in the output I from the photocell amplifier PCA at the instant when a pulse of the waveform CL arrives at the left deflection gate LDG. This gate is thus opened to allow the passage of this pulse in the train CL which is then operative to apply a left deflection 'voltage to one of the X-plates of the cathode ray tube CRT. Similarly, should the s'canningbeam SB strike a right deflection marking b in Figure 5, a pulse of the train Cs is gated-by the right deflection gate to apply a right deflection voltage to the other X-plate of the cathode ray tube'CRT. The scanning .beam SE is thus deflected laterally for a distance equal to one half the width of acolumn cso that it correctly traverses the markings mo, m1.

It will be understood that separate pulse train generators may be used to generate the respectivewaveforms'Ct, CR, On, the separate generators being kept in step by any known method.

Figure 15 illustrates a's'imple citcuit'sirrange ment for providing a digital output corresponding 'to a function'of two independent variables. In this arrangement, the movable pattern ME carries markings corresponding to successive values of a function of one variable-say J(.r) and is scanned in themanner described with reference to Figure 11.

The Y-plates of the cathode ray tube CRT are ,fed with a time-base voltage from the time-base circuit TB in the normal manner whilst theX- plates are connected to another circuit (not shown) at the terminals 8 for producing voltages corresponding to values of a function of an independent variable Y. In the absence of any voltage applied to the'terminals 8, the circuit operates to' produce an output waveform D from the digit gate DG corresponding to the function represented by the pattern ME. If, however, 9. voltage is now applied at the terminals 8, th scanning beam SB will be deflected relative to the pattern'ME so that the output D is varied in accordance with the function j r+y 'Ihe circuit may, with advantage, be used where 11 represents small corrections to the variable :17, although it will be understood that it may be used in all cases where it is required that the digital output D shall represent a function of the algebraic sum or two independent variables.

Any of the forms of circuit described above for the prevention of transmission of a wrong nu ber may be 'adoptedin the circuit "shown in Figure 15.

It will .be understood that in the case of a pattern representing a function of a variable it may be arranged that some numbers are omitted over some parts of the pattern or some numbers are repeated in consecutive columns 0. When, however, the means described in connection with Figs. 5 to for avoiding wrong numbers are used, it is necessary that adjacent coluxims'shofild represent numbers difiering by 0 or 1. This can be arranged by providing suflicient steps of the variable in relation to the number of digits representing the function. Alternatively, if it is not convenient to provide this sufiiciency of steps of the variable, or if the function is discontinuous, so that some adjacent columns 0 represent numbers differing by more than one, the initiator marking 2 of Fig. 7 may be reduced in width by a factor of two and twice the number of such markings 2 may be used, the centre-line of each marki 2 being coincident with a edge .e

umn c and all such edges having markings 2 cpposite them. In scannin the patterns of Figs. 1, la and 3, if the scanning spotsize is small compared with the width of the digit markings mo, m1 and if the scanning beam SBdoes not deviate appreciably from its correct direction, no wrong numbers will arise and it will not be necessary to adopt the correction means described in connection with Figs. 5 to 10, nor to arrange for adjacent columns to represent" numbers differing by no more than unity. Such means may, however, be employed as a precaution, ifthe spot size is small, and will be'desirable if the spot size is compara-- ble with the width of the markings.

Fig. 16 shows diagrammatically a possible form of markings for deriving digital'data in the ternary scale or scale of three. In this scale, there are three digits, 0, 1, 2, each of which can be represented by a corresponding marking mo, m1, m2 having a distinctive characteristic representing its respective digit value. Thus, as shown in the figure, the movable pattern ME is of a generally light-absorbent nature, and is formed with transparent markings m1 and reflecting markings m2 representing, say the digit values 1 and 2. Further markings mo of a light-absorbent nature are formed to represent, say, the digit value 0. The movable pattern ME is illuminated by a scanning beam lamp SL, and photocells PC1 and PC2 'are located on opposite sides of the pattern. The photocell PC1 receives light transmitted through the transparent markings m1 while the photocell PC: receives light reflected from the reflecting markings m2. The output pulses from the'photocell PC1 represent the digit of value 1 whilst the output pulses from the photocell PCz represent the digits of value 2 in the pattern. For ease of illustration, the scanning beam lamp SL is shown in three alternative positions to indicate relative movement between the said lamp and the pattern ME. It will be understood that only one lamp SL would be used in a series system to illuminate both photocells P01 and P Cz whilst the pattern ME would be traversed in the appropriate direction with respect thereto. In a parallel system employing three different types of markings mo, m1, m2, there would be a lamp SL and a'slit to illuminate the full height of the pattern and as many of each of the photocells PCi and 1202 as there are digits in the pattern.

The arrangement of Figure 1 6 may alternatively be used as a binary digital system in which control markings-for example, the markings ma -for indicatin instructions (suchas beam shift) to the apparatus are of a distinctive nature from markings mu, mi representing digits in the pattern. Alternatively, markings m1 and 1112 can be used to distinguish both binary digit values 0 and 1 from the background of the pattern.

Details of the circuits represented as block diagrams in certain of the preceding figures will now be described, these circuits merely being examples chosen from the many known arrangements which will perform the required functions as will be obvious to those skilled in the art, and 2101', in themselves constituting part of this invenion.

The photo-cell amplifier represented by the PCA block is shown in Fig. 17 as a direct coupled pentode amplifier, the pentode having its control grid 56 fed with the output from the photocell and the output from this amplifier being taken from the anode at 51.

e digit Pi s? $9426 ep fi t d b he W9 which are connected through respective condensers 68 and GI to the suppressor and control grids 62 and 83 respectively of a short suppressor base pentode 64. Each of these grids is negatively biased, through diode D. C. restoring circuits indicated at 65 and 86, respectively, so that anode current can only fiow when both grids 62 and 63 are simultaneously driven positive relativ to their quiescent potentials.

The output from the pentode 64 is taken from its anode 61 through a condenser 68 to the control grid 88 of an inverter triode I8, a further diode D. C. restoring circuit H being connected between the grid 68 and H. T. negative. The final output of the gate is taken from the anode I2 of the inverter triode I8.

at its anode I! from which the output of the gate is derived. The gate thus operates to deliver a pulse to the output circuit indicated at D in Figs. 11, 13, 14, and 15 only when a pulse from the photo-cell amplifier coincides in time with a clock pulse from the pulse train generator. Consequently, the output D in the figures referred to represents the wanted signals from the element ME carrying the markings being scanned.

Fig. 19 shows a typical circuit for use as the pulse train generator represented by the block PTG in Figs. 11 and 15, consisting of a free-running multivibrator I3 in the form of a double triode having anodes I4 and I5 interconnected in the conventional manner. The output from the multivibrator I3 is taken off the anode 14, at

- 16, the said output consisting of a train of substantially square-shaped positive-going pulses as indicated.

In order that a trigger pulse for operating the time base circuit (yet to b described in detail) may be derived from this train of pulses the anode I4 is also coupled through a condenser 11 and resistance I8 to the first of a series of frequency divider double triodes 18, 8| and 83, it being understood that many more than the three stages shown are normally required.

The double triode 19, as is known, when supplied with the pulse train shown at IE, will produce at its one anode 88 a secondary train of pulses having a repetition frequency equal to one half of the repetition frequency of the pulses in the input pulse train.

Similarly, the output from the anode 82 of the second frequency divider stage is a train of pulses having a repetition frequency equal to one half of that of the input derived from the anode 88 of the first frequency divider I9 and the output from the anode 84 of the third frequency divider stage 83 is a train of pulses having a repetition frequency equal to one half of that of the input derived from the anode 82 of the second stage 8|. The anode 84 of the final frequency divider stage 83 is connected through a diode 85 to an output channel 86. The output at 86 consists of a train of peaked negative trigger pulses having a repetition frequency equal to V 3 of that of the clock pulses produced by the multivibrator 13. Any number of frequency divider stages may be used, the final output having a pulse repetition frequency which is &1; of that of the clock pulse train shown at I6, where n is the number of frequency divider stages fed from the multivibrator 13.

Fig. 20 shows an extension of the circuit of Fig. 19 which renders it suitable for use as the pulse train generator represented by the block PTG in Fig. 13 where it is required to have a number of outputs in each of which pulses will occur at definite time intervals in relation to the final frequency divided output derived from the circuit illustrated in Fig. 19. In Fig. 20 the output 88 from the pulse train generator PTG of Fig. 19 is inverted and squared by a combination comprising the triode 81 and diode 81a before being fed by Way of a cathode follower 88 into a delay chain composed of series-connected delay lines 881, 892 8811. Each delay line, such as 881, is operative to delay the transmission of a pulse therethrough by one digit time, i. e. the period of one cycle of the output derived at 18 from the pulse train generator PTG of Fig. 19. The output from each delay line 891 88" is fed as one of the inputs to a separate digit gate DG which is arranged as shown in Fig. 18. The other input to the digit gate DG is composed of the train of clock pulses derived from the output I8 of the pulse train generator PTG and, hence, the output from each digit gate DG represents a series of rectangular pulses bearing a selected fixed time relationship to the pattern of markings bein scanned.

Fig. 21 illustrates a time base circuit such as is represented by the block TB in Figs. 11, 13, 14, and 15. This circuit comprises a double triode 88 arranged to function in the known manner as a one shot flip-flop. A negative-going trigger pulse, indicated at 9I, derived from the frequency divided output 86 of the clock pulse generator shown in Fig. 19 is applied to the input terminal 92 of the circuit and is fed through a diode 83' to one grid 94 of the flip-flop 88 to cause the execution of one complete cycl of anode voltage changes therein whereby a positive-going rectangular pulse is produced at the anode 95. This pulse is fed as a gating wave, through a. condenser 96, to the suppressor grid 8! of a Miller integrating pentode 98 the control grid of which is connected in the normal fashion, through a grid leak, to a supply of positive potential marked 88. An output is derived from the anode I88 in the form of a saw-tooth wave form as indicated at I8I. This output has a frequency which is controlled by the repetition frequency of the input trigger pulses 9|. The speed of linear rundown of the anode potential of the pentode 98, durin the period of application of the gating wave, may be varied by adjusting the potential of the point 89 to which the grid leak is returned.

Fig. 22 shows a suitable circuit arrangement for use as the beam deflector circuit represented by the block BD in Fig. 13. The input channel I82 receives an output pulse from the initiator gate represented by the block IG in Fig. 13 and this latter gate comprises a digit gate circuit as shown in Fig. 18 in which one input 58.0r 58, is the output from the photo-cell amplifier PCA, as before, whereas the other input is the output from an appropriate digit gate DG of the circuit shown in Fig. 20 (instead of the clock-pulse train). It will be appreciated that the initiator gate will deliver a pulse to the input channel I82 conditions are obtained.

1, 9! Pic- 22 on y when there is coincidence in time :a pulse r m th photocell amplifier and a mal eirom the elected digit sa of 20. The pulse tram the initiator gate IG operates to set adirect-coupled flip-flop I03 which is n turn mtby an appropriate pulse fed to the input 4 during the next minor cycle period. When the .flmflop 1.03 is s t. the cmrent thr ugh an auxtriode I05 is varied, due to the connection :Irom the anode Hi6 of the flip-flop to the grid 19] of the :triode, to a predetermined extent suffieientto vary potential on its anode I08 .by an wh ch will, when applied to the appro hri te deflecting plate of the cathode ray tube sh wn a GRTin F g. 3. deflect the beamin this K tube laterally throu h a distance approximately equalto one spot .diameterlon the screen thereof.

Fig. 23 shows a circuit diagram which is suitable ior use in both the blocks TBCI and TBC2 Fig. 1.3, the input and output connectionsbesuitably adjustedin either case, as will be indleated. Considering first the circuit as used at a clock pulse (In derived from an appropriately selected digit gate DG of Fig. 20, this being represented on the wave form .Cr in Fi 10, is fed to one input channel I09. The

other input channel I I0 receives the output from the photo-cell amplifier of PCA of Fig. 13 (shown in detail in Fig. 17) The input Cu is differentiated by the capacity resistance network III,

.. I 12 to produce a short positive-going pulse corresponding with the-start of the pulse Cu and a negative-going pulse corresponding with the end of the pulse Cu. These pulses are simultaneously applied, in each case, to the one cathode I I3 of a double diod II and to one anode I of a similar double diode I I6. At the same time, the input from the photo-cell amplifier PCA on the channel H0 is similarly differentiated by the capacity resistance network III, I I8, and the resultant positive-going and negativevgoing pulses are applied to both anodes I IQ of the doubl d ode H4 and to both cathodes I20 of the double diode I I6.

If the output from the photo-cell amplifier ,PQA of Fig. 13 is early in relation to the pulse the negative-going pulse resulting from the differentiation of this output coincides in time with the negative-going pulse derived from the differentiation of the input Cu so that the negative output from the double diode I I6 is of greater magnitude than the positive output from the double diode I I4. On the other hand, if theoutput from the photo-cell amplifier PCA of Fig. 13 5 late in relation to the pulse Cu, the reverse If the output from the photo-cell amplifier PCA occurs in the c rrect timed relation to the pulse Cu, the positive-going and negative-going differentiated pulses resulting from that output will be symmetrically disposed in time about the corresponding differentiat d pulses derived from the pulse Ca.

The positive and negative outputs from the I, IIS are mixed, amplified by a pentode I2I and then separated and lengthened by the circuits associated with a triode I22 and double diode I23. The mixed output from the double diode I23 is then integrated by the M l er feed-back circuit associated with the double triode amplifier I24 and the resultant output potential at the anode I25 of this double triode is applied to the beam deflecting plates of the cathode ray tube CRT to displace the trace in the direction of scan.

When this circuit is used at TBC2 in Fig. 13,

the input I09 receives the clock pulse Co on the waveform Cr in Fig. 9,:instead of the pulse Gtl as Just described, and the input I10 receivw the output from the photo-cell amplifier PCA as before. The components of the circuit functionas already described but the output taken. oil the anode I25 of the double triode I24 is utilized to govern the potential of the adjustable positive supply at 99 in the time base circuit TB shown in detail in Fig. 21. Thus the scanning sweep of the beam in the cathode ray tube CRT may be speeded up or slowed down as required.

Fig. 24 shows the circuit for use where indicated by the block PTG in Fig. 1.4, this including -,a circuit such as illustrated in detail in Fig. 19 and represented in 24 by the small rectangle marked PTG (Fig. 19) ,the outputs I6 and 86 for clock pulse trains and trigger pulse trains, respectively, being indicated. From the output .15 the clock pulses are supplied by way of a difterentia-ting circuit I26, I21 and a diode limiter I28 to an inverter trlode I29 which feeds, through .a cathode follower I30, the resulting very short pulses into delay line sections I3I, I3Iz I3Ix, I3-In connected in series. Each delay line section produces a delay in the transmission of a pulse therethrough which is very much shorter than the delay produced by the delay line sections in Fig. 20. Since the pulses .fed into the delay line in the present case are extremely short the required pulse trains -CL and Ca (refer also to Fig. 6) may be taken off at the appropriate sections (shown as the sections I 3Ix and I3In) without further provision.

The circuit represented by the block LDG in Fig. 14 comprises a combination of the circuits shown in detail in Figs. 18 and 22, the beam de fiection circuit being connected to follow the digit pulse gate. In this case the output of the beam deflection circuit will control the appropriate deflection plate in the cathode ray tube CRT to produce a lateral deflection of the scanning beam in the appropriate direction. Similarly, the circuit represented by the block LDG in Fig. 14 also comprises a digit pulse gate as illustrated in Fig. 18 followed by a beam deflection circuit as illustrated in Fig. 22. In this case, however, the output of the beam deflection circuit controls the opposite deflection plate in the cathode ray tube CRT so that the scanning beam will be shifted laterally in the opposite sense.

What we claim is:

1. In a digital data transmission system, a free- 1y movable element carrying markings, indicative of its position relative to a datum position according to a predetermined mathematical function, said markings being arranged in columns located side by side in the direction of motion of the element and the markings in each column representing in the binary system of numbers the numerical value of the function at that particular point on the element, a relatively fixed scanning device for continuously scanning the markings in the direction. of length of the columns at a given location, means for producing a train of pulses corresponding to the binary number represented by the markings in the column being a clock pulse generator, means for feedcorresponding to the binary number of the col-- umn being scanned to the said gating circuit to control the gating operation thereof, and means for transmitting the trains of gated clock pulses to represent binary values of th mathematical function.

2. In a digital data transmission system, the combination of a freely movable element carrying code markings indicative of its position relative to a datum position according to a predetermined mathematical function, said code markings being arranged in columns located side by said in the direction of motion of said element and the code markings in each column representing the numerical value of the function at the related location on the element in the binary system of numbers, a relatively fixed scanning device for continuously scanning the code markings in the direction of the length of the columns at a predetermined scanning location, means responding to said scanning device for producing a train of pulses corresponding to the binary number represented by the code markings in the column being scanned, a clock pulse generator, means for feeding selected regularly recurrent pulses from said generator to said scanning device for determining the instants of commencement of successiv scanning operations, means for feeding all of the pulses from said clock pulse generator to a gating circuit, means for feeding the train of pulses corresponding to the binary number of the column being scanned to said gating circuit to control the gating operation of the latter, means for transmitting trains of gated clock pulses from said gating circuit to represent binary values of the mathematical function, initiating markings on said movable element tor at least certain of said columns of code markings and displaced in the direction of movement of said element relative to the respective column of code markings to be scanned by said scanning device in the event that the latter is not accurately aligned with a column or code markings, and means responding to scanning of said initiator markings to eiiect deflection of said scanning device in the direction of movement of said element into alignment with the column of code markings disposed at the scanning location.

3. In a digital data transmission system; the combination according to claim 2, wherein said means responding to scanning of said initiator markings to deflect the scanning device comprises an initiator gate circuit, means for feeding selected regularly recurrent pulses to said initiator gate circuit at intervals corresponding to the location of said initiator markings on said element in the direction of said columns, means for feeding pulses corresponding to the initiator markings to said initiator gate circuit for controlling the gating operation of th latter, means for deflecting said scanning device, and means for feeding the gated pulses issuing from said initiator gating circuit to said deflecting means for controlling the operation of the latter.

4. In a digital data transmission system; the combination according to claim 2, wherein said columns of code markings are spaced apart in the direction of movement of said element, and said initiator markings for said certain columns are disposed at opposite sides of the respective column and displaced relative to each other in the direction of the length of the column.

5. In a digital data transmission system; the combination according to claim 4, wherein said means responding to scanning of said initiator markings to deflect the scanning device comprises first and second initiator gate circuits, means for feeding selected regularly recurrent pulses to said initiator gate circuits at out of phase intervals corresponding to the locations of said initiator markings at the opposite sides of the columns of code markings, means for feeding pulses corresponding to the scanned initiator markings to said initiator gate circuits for selectively controlling the gate operation of the latter, means for deflecting said scanning device in opposed directions, and means for feeding the gated pulses issuing from the selected one of said first and second initiator gate circuits to said deflecting means for controlling the operation of the latter in the corresponding direction.

6. In a digital data transmission system; the combination according to claim 2, further includ-' ing timing markings on said element at the opposite ends of said columns of code markings to be scanned by said scanning device at the commencement and end of each scanning sweep, means for producing timing pulses corresponding to the scanning of said timing markings, means synchronized with said generator for producing regularly recurrent control pulses which normal- 1y are in a predetermined out of phase relationship with respect to said timing pulses and of the same frequency, and means responding to deviations from said predetermined out of phase rela-.

tionship betweensaid control pulses and timing pulses and to difierences in the frequencies thereof for correcting the timing and sweeping speed of said scanning device.

7. In a digital data transmission system; the combination according to claim 2, wherein said columns of code markings are contiguous to each other, and said initiator markings are provided only for the columns having code markings representing odd numbers in the binary system of numbers, said initiator markings being disposed adjacent the ends of the respective columns of code markings at which the scanning of said columns begins and having their centers substantially aligned with side edges of the code markings in the related columns.

8. In a digital data transmission system; the combination according to claim 7, wherein said means responding to scanning of said initiatormarkings to deflect the scanning device comprises an initiator gate circuit, means for feeding selected regularly recurrent pulses to said initiator gate circuit at intervals corresponding to the location of said initiator markings on said element in the direction of said columns, means for feeding pulses corresponding to the initiator markings to said initiator gate circuit when said scanning device scans said initiator for controlling the gating operation of the initiator gate circuit, means for deflecting said scanning device in the direction from said side edges toward the centers of said code markings, and means for feeding the gated pulses issuing from said initiator gate circuit to said deflecting means for controlling the operation of the latter.

9. In a digital data transmission system, the combination of a freely movable element carrying code markings indicative of its position relative to a datum position according to a predetermined mathematical function, said code markings being arranged in columns located side by side in the direction of motion of said element and the code markings in each column representing the numerical value of the function at the related location on the element in the binary system of numbers, a relatively fixed scanning device for continuously scanning the code markingsinthefiiredtiomdf the length'of'the'columns at a prede'terminedscanning" location, for producing 3, "trainor pulses corresponding' 'to" the binary "number represented'by thecodemarkings in the 'column'being scanned, a clock pulse 'gen- 'erator, -means *ior feeding *selected regularly recurrent pulses from said generator i to said scanning device ier-effecting the suc'cessive I scanning operations of the latter, a gating circuitpm'eans "for feeding all 'of=the'pulses from said clock-pulse generator 7 to said gating circuit, means for' feeding' thetrain of. pulses corresponding to the binary number of the column' being scanned to said gating circuit to'controlgatmgoperation of the latter, means for transmitting trains -of 'gated clock pulses from" said gating circuit to represent binary va1ues-ofthemathematical iunction,='timing markingsonsaid'element atthe'oppositeends 'of'said-columns of code markingsto -be*scanned by said scanning device at the "commencement *1 *andend of each scanning sweep, means foriproducing "timing" pulses corresponding to the scanning of 'said timing =mai'kings, -means synchronizecl' with said generator-Tor producing regularly recurrentcontror pulses -which=normally are in a 'predetenninedoutf phaserelation'ship with respectto said"timing pulses and of the samefrequency;and-means'responsive' to deviations 'from -said predetermined outof phase relationship-be- -tween 'saidcontrol' andtimingpiilses and to ditinthe frequenciesthereof for shiftin the phase and correctingthespeedcf'scanning of saidscanning device.

-10. -=In a digital data *transmissionsystem; the 'contibination according to claim 9, wherein -sai'd timing pulse producing means is "operative "to generate *timing pulses each "of which includes an -initial positive ?portion and a following neg-- ative portion and-'said control'pulse producing means isoperativetirgenemte eontrokpli-lsesmh of-which-includes an initial negative portion and *afollowing positive portion with the negativeand positive portions of said control pulsesnormaliy "coinciding in tim with the-positive and negative portions,respectively/of the related timing-pulses "to -'sub'stantia1ly cancel each other, and' whenein said "nhase shifting *and speed correcting means is operative to shift "the phase and eorrect the speed *of scanning of i the-scanning *defiiee in ene direction when the positive -'po1"-tions 6f related control and timing pulses ove'rlap and to shift the phase and correct the -speed of scannin 1-1 in "the oppositedireetion when the negative portions of-related control an'd timing' pul-ses overlap.

- J COALE'S.

"SIDNEY NORMAN DOUGLAS HIBL.

SAMUEL EDWIN HERS QM.

JOHN TINDALE.

l-Re'fe'rences Cited in the 18 idf lthis patent "UNITED "STATES PA'I'ENTS Number Name Date 464,513 Molyneux Dec. .8, 1,891 v2,057,773 Finch Ot.-20,l193 6 2,302,009 Dinkinson Novvl'l, 1942 -2-,318,591 Couffignal. v May .1 1, i 1943 2,348,535 Goodale i iMay 9, 1944 2,397,604 Hartley et .al. Apr. 2,19% 2,469,229 Smith, J23, ve1; a1.. Oct..;15,. 1'9'46 2,410,265 Broido' Dots-2'9, 1946 2,428,990 Rajchman. Oct .14, .1947 2,429,228 'Herbst fonts--21, I947 2,431,591 Synderet a1 Nave-25,1947 2,473,691 .Meacham. June'21, 1949 2,534,369 Ress .Dec. ,19, 19

Images